對話式圖片編輯
描述你真正需要的照片編輯
用日常語言規劃照片編修。先定義要改什麼、哪些內容要保留，再把編修結果和原圖比較。
模型
這個創作工具只會生成靜態圖片。漫畫與影片請接續使用連結工具；研究策略不是可選的正式生產模式。
圖片記錄
Google Vertex AI · gemini-3.1-flash-image · · 1280 × 859 px
原始生成提示詞
Use case: photorealistic-natural / product-mockup. Create a natural window-lit editorial PHOTOGRAPH of exactly TWO anatomically normal adult hands holding one landscape matte comic proof sheet above a real warm oak desk. Image1 supplied is the original adult botanist and mint robot reference sheet; image2 is the COMPLETE original SIX-PANEL greenhouse comic. The held paper must have landscape 3:2 proportions so Image2 fits WITHOUT rearrangement. Put the exact complete Image2 sequence on the held sheet as THREE columns by TWO rows, with a clean ivory outer margin: upper row arrival, pot on bench, gloved seed close-up; lower row watering, seedling reaction, final greenhouse view. Preserve all six distinct scenes in that order; DO NOT repeat watering, remove seed, reformat into two columns, add text or combine panels. Keep the original adult's teal jacket, coral scarf, dark curly hair, sand trousers and original mint round robot. Hold the paper by its two lower blank outside corners only; all artwork corners remain visible and hands cover no panels. The source character-reference sheet from Image1 rests flat farther back on the desk, partly visible beside the held print. Only the printed artwork is ink/watercolor illustration; hands, table, thin paper grain and room are a convincing photograph. No visible faces, booklet, screen, logos, titles, numbers, writing, fake UI, metrics or product packaging. Natural skin/five fingers, coherent optical perspective, subtle paper flex, quiet soft daylight. The completed hand-held story takes 75 percent of frame. Fictional lifestyle presentation only, never a customer photograph or research result. Landscape 3:2.
查看完整作品規劃具體結果
不同外套，同一位原始角色
規劃一個示範性編修，將原始成年角色的外套顏色改掉，同時保留臉、姿勢和圍巾。可作為有明確對照的控制任務。
準備內容
- 經授權的原創圖片。
- 簡短的修改簡報，以及需要保持穩定的特徵清單。
圖片記錄
Google Vertex AI · gemini-3.1-flash-image · · 1280 × 859 px
原始生成提示詞
Use case: illustration-story. Make ONE original finished comic page in a clean three-column by two-row grid, exactly SIX equally sized panels, narrow even ivory gutters and complete outer margins. The provided image is the character identity and wardrobe reference, not a page to copy. Preserve its adult botanist with warm brown skin, short dark curly hair, teal jacket, coral scarf, sand trousers, brown boots, and round mint helper robot with two dark eyes and exactly one short antenna. Quiet elegant ink-and-watercolor comic style, teal/coral/sage/ochre palette. Tell a very simple readable rooftop greenhouse planting story: panel1 wide view botanist and robot arrive at the greenhouse carrying a small terra-cotta pot; panel2 medium view the botanist places the pot on the wooden bench; panel3 close view gloved adult hands holding a seed above the pot; panel4 medium view botanist and robot water the pot with a small cream watering can; panel5 close view botanist and robot look at the tiny seedling; panel6 wide view both stand by the finished row of three potted seedlings in afternoon light. All people are adults; each scene has one botanist and one robot unless only hands are framed. The pot remains terra-cotta, the greenhouse stays structurally coherent, identity and outfit remain recognizable. No writing, numbers, logos, speech balloons, captions, lettering, illegible marks, third-party characters, UI, photorealistic environment or claims of tested consistency. Landscape 3:2.
查看完整作品先限制第一個變更
修改一個屬性，才能看出是否有照指示執行。
保護身份與幾何結構
將臉、手、輪廓和物件形狀與原圖比對。
檢查整張圖
如果背景或角色不小心變了，光是外套顏色正確還不夠。
有價值的編輯，會同時看它是否完成需求，以及是否保留了重要細節。
實用的起點
從有授權的來源開始
選擇一張你有權編輯的清晰圖片，並保留原圖方便比較。
描述一個可控變更
請明確指出要改什麼，以及哪些細節應保持穩定，例如人物身份、物件形狀或構圖。
比較編輯區內外
檢查請求區域與周邊細節。在保留結果前，先看光線、邊緣、人體結構與是否有非預期變化。
檢查重點
圖片記錄
Google Vertex AI · gemini-3.1-flash-image · · 1280 × 859 px
原始生成提示詞
Use case: photorealistic-natural / product-mockup. Create a natural window-lit editorial PHOTOGRAPH of exactly TWO anatomically normal adult hands holding one landscape matte comic proof sheet above a real warm oak desk. Image1 supplied is the original adult botanist and mint robot reference sheet; image2 is the COMPLETE original SIX-PANEL greenhouse comic. The held paper must have landscape 3:2 proportions so Image2 fits WITHOUT rearrangement. Put the exact complete Image2 sequence on the held sheet as THREE columns by TWO rows, with a clean ivory outer margin: upper row arrival, pot on bench, gloved seed close-up; lower row watering, seedling reaction, final greenhouse view. Preserve all six distinct scenes in that order; DO NOT repeat watering, remove seed, reformat into two columns, add text or combine panels. Keep the original adult's teal jacket, coral scarf, dark curly hair, sand trousers and original mint round robot. Hold the paper by its two lower blank outside corners only; all artwork corners remain visible and hands cover no panels. The source character-reference sheet from Image1 rests flat farther back on the desk, partly visible beside the held print. Only the printed artwork is ink/watercolor illustration; hands, table, thin paper grain and room are a convincing photograph. No visible faces, booklet, screen, logos, titles, numbers, writing, fake UI, metrics or product packaging. Natural skin/five fingers, coherent optical perspective, subtle paper flex, quiet soft daylight. The completed hand-held story takes 75 percent of frame. Fictional lifestyle presentation only, never a customer photograph or research result. Landscape 3:2.
查看完整作品研究方向
工作流程背後的研究
衡量非預期變化
我們正在探索區域修正方法，並明確審核被編輯區域與其外部場景。
同時檢查保留度與指令遵循
編輯可能在照做一項指示時，卻改掉另一個重要細節。審核時應把兩種結果都記錄下來。
研究方向
視覺審核流程
哪些檢查能及早抓出身份或故事錯誤，哪些則需要更仔細看？
我們會怎麼評估: 追蹤漏掉的重大錯誤與誤判被拒的分鏡，並同時記錄審核成本與延遲。
研究方向
定向角色修復
能不能只修正一個錯誤的角色細節，同時保留已通過的場景？
我們會怎麼評估: 每次提出修正後，都要檢查編輯區、周圍像素和故事內容。
了解界線
- 精準保留與區域控制取決於所選的編輯工具。
- 這個規劃案例不是有量測的前後對照產品測試。
這些問題參考了第一手研究；但不能據此認定 CyaniModel 已有對應結果。
常見問題
我可以用日常語言編輯嗎？
在支援的圖片編輯工作流程中，清楚描述預期變化。開始前先確認可用的參考與編輯選項。
怎麼避免改到照片其他部分？
先說明要保留的細節，做一次可控的初步編修，然後比對整體結果。只有文字指示並不能保證像素完全不變。
我可以編輯真實人物的照片嗎？
請使用你有權編輯與分享的素材。仔細檢查身份、情境與預定用途；此處示例使用的是原創虛構角色。
繼續你的故事
把預期變更和需保留的細節說清楚。
繼續前往 LlamaGen，在生成前查看可用工具、模型選項、帳號需求與積分。